Figures for MS: Spatiotemporal variation in mountain stream metabolism and nitrogen cycling across contrasting flow regimes
Kelly Loria
2025-02-26
Research Questions:
How does stream metabolism vary based on catchment characteristics, antecedent flow conditions, and nitrogen availability?
- Figures 2, 3
Is variance in nitrogen demand and instream uptake (either as ammonium or nitrate) more related to biological (GPP, ER, biomass or chlorophyll-a), physical processes (flow and water temperature), or ambient nutrient and organic matter availability?
- Figures 4, 5
How does in-stream productivity and nitrogen supply and demand dynamics change within stream catchments and across dry (2021 and 2022) and wet water years (2023)?
- Figures 6
Figure 1: Map
Not included here.
Pause to make a column for biomass to chla ratios And also make biomass in in micro grams to more comparable to chl-a
Question 1
How does stream metabolism vary based on catchment characteristics, antecedent flow conditions, and nitrogen availability?
Figure 2: Metabolism Time series
- The magnitude of GPP and ER across all sites
Where GPP is in green, ER is in purple, NEP is in black dataframe:
metab_dat
Time series metabolism plots for each site
Descriptive stats to describe stream based differences in
Both BW to GB - large to small streams, as well as variation with stream.
Water temp
Flow
OM
Biomass
Chla
GPP
|ER|
BW catchment:
## mean_AFDM_mgg min_AFDM_mgg max_AFDM_mgg mean_Chla_ugL_Q min_Chla_ugL_Q
## 1 16.642 8.567217 49.13182 45.981 0.3621839
## max_Chla_ugL_Q mean_AFDM_ugcm2 min_AFDM_ugcm2 max_AFDM_ugcm2 mean_GPP min_GPP
## 1 417.6117 51.968 0 369 1.883 0.001
## max_GPP mean_ER min_ER max_ER mean_wtr_m min_wtr_m max_wtr_m mean_Q_m
## 1 11.54822 -12.379 -28.72293 -1.725866 8.614 0.2916944 20.4719 0.715
## min_Q_m max_Q_m
## 1 0.009937639 8.531119## # A tibble: 2 × 16
## substrate mean_PO4_ugL_dl min_PO4_ugL_dl max_PO4_ugL_dl mean_NO3_mgL_dl
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 pw 10.5 4.26 62.3 0.063
## 2 sw 9.65 1.68 26.2 0.015
## # ℹ 11 more variables: min_NO3_mgL_dl <dbl>, max_NO3_mgL_dl <dbl>,
## # mean_NH4_mgL_dl <dbl>, min_NH4_mgL_dl <dbl>, max_NH4_mgL_dl <dbl>,
## # mean_pH <dbl>, min_pH <dbl>, max_pH <dbl>, mean_DOC_mgL_dl <dbl>,
## # min_DOC_mgL_dl <dbl>, max_DOC_mgL_dl <dbl>GB Catchment
## mean_AFDM_mgg min_AFDM_mgg max_AFDM_mgg mean_Chla_ugL_Q min_Chla_ugL_Q
## 1 22.175 8.710089 58.84203 48.077 2.046401
## max_Chla_ugL_Q mean_AFDM_ugcm2 min_AFDM_ugcm2 max_AFDM_ugcm2 mean_GPP
## 1 260.7374 86.347 0 292 0.107
## min_GPP max_GPP mean_ER min_ER max_ER mean_wtr_m min_wtr_m
## 1 4.520732e-05 3.150141 -4.97 -19.23026 -0.4027769 7.826 0.01800694
## max_wtr_m mean_Q_m min_Q_m max_Q_m
## 1 15.66177 0.054 0.0009168555 0.9424983## # A tibble: 2 × 16
## substrate mean_PO4_ugL_dl min_PO4_ugL_dl max_PO4_ugL_dl mean_NO3_mgL_dl
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 pw 26.7 8.5 109. 0.084
## 2 sw 14.1 8.47 25.3 0.017
## # ℹ 11 more variables: min_NO3_mgL_dl <dbl>, max_NO3_mgL_dl <dbl>,
## # mean_NH4_mgL_dl <dbl>, min_NH4_mgL_dl <dbl>, max_NH4_mgL_dl <dbl>,
## # mean_pH <dbl>, min_pH <dbl>, max_pH <dbl>, mean_DOC_mgL_dl <dbl>,
## # min_DOC_mgL_dl <dbl>, max_DOC_mgL_dl <dbl>Break out by stream reach location:
BWL:
BWU:
GBL:
GBU:
Kernal density plots of GPP and ER
Colored by years
Modified figure 2
(Q1) - How does stream metabolism vary based on catchment characteristics, antecedent flow conditions, and nitrogen availability?
General drivers of productivity
Both standing stocks
A. Epilithic biomass
B. Epilithic chl-a
Ecosystem fluxes
C. Gross primary productivity (GPP)
E. Absolute value of ecosystem respiration (ER)
Each productivity response is individually modeled within stream
using glm
dataframe = covariat_datq
A. Biomass models
BW:
biomass_mod_bw <- lmer(log(AFDM_ugcm2+1)~
scale(NO3_mgL_dl* 1000) +
scale(NH4_mgL_dl* 1000) +
scale(PO4_ugL_dl) +
scale(AFDM_mgg)+
# scale(DOC_mgL_dl)+
scale(wtr_m) +
scale(Q_m) + (1|site), data=covariat_datq_BW)
summary(biomass_mod_bw)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(AFDM_ugcm2 + 1) ~ scale(NO3_mgL_dl * 1000) + scale(NH4_mgL_dl *
## 1000) + scale(PO4_ugL_dl) + scale(AFDM_mgg) + scale(wtr_m) +
## scale(Q_m) + (1 | site)
## Data: covariat_datq_BW
##
## REML criterion at convergence: 231.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.14121 -0.25057 0.02282 0.43491 2.21972
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.07452 0.273
## Residual 1.81282 1.346
## Number of obs: 66, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.46061 0.29677 0.97047 11.661 0.05825 .
## scale(NO3_mgL_dl * 1000) 0.23732 0.15059 58.86194 1.576 0.12040
## scale(NH4_mgL_dl * 1000) -0.19969 0.18151 58.07195 -1.100 0.27581
## scale(PO4_ugL_dl) 0.08077 0.12331 58.59587 0.655 0.51500
## scale(AFDM_mgg) 0.18559 0.29544 58.06074 0.628 0.53235
## scale(wtr_m) 0.68938 0.25854 58.00248 2.666 0.00992 **
## scale(Q_m) -0.07222 0.15056 58.94510 -0.480 0.63326
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(NO*1 s(NH*1 s(PO4_ s(AFDM scl(_)
## s(NO3_L_*10 0.022
## s(NH4_L_*10 -0.119 -0.356
## scl(PO4_L_) 0.012 -0.131 -0.045
## scl(AFDM_m) -0.068 -0.024 -0.071 0.177
## scal(wtr_m) 0.290 0.229 0.074 -0.027 -0.169
## scale(Q_m) -0.159 0.196 0.243 -0.253 -0.201 0.410## scale(NO3_mgL_dl * 1000) scale(NH4_mgL_dl * 1000) scale(PO4_ugL_dl)
## 1.321235 1.298738 1.112896
## scale(AFDM_mgg) scale(wtr_m) scale(Q_m)
## 1.076703 1.269469 1.435108## R2m R2c
## [1,] 0.1554381 0.1887849GB:
biomass_mod_gb <- lmer(log(AFDM_ugcm2+1)~
scale(NO3_mgL_dl* 1000) +
scale(NH4_mgL_dl* 1000) +
scale(PO4_ugL_dl) +
scale(AFDM_mgg)+
#scale(DOC_mgL_dl)+
scale(wtr_m) +
scale(Q_m)+(1|site), data=covariat_datq_GB)
summary(biomass_mod_gb)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(AFDM_ugcm2 + 1) ~ scale(NO3_mgL_dl * 1000) + scale(NH4_mgL_dl *
## 1000) + scale(PO4_ugL_dl) + scale(AFDM_mgg) + scale(wtr_m) +
## scale(Q_m) + (1 | site)
## Data: covariat_datq_GB
##
## REML criterion at convergence: 207.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.95119 -0.45345 0.01751 0.57955 1.76719
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.000 0.000
## Residual 1.592 1.262
## Number of obs: 61, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.70075 0.20246 54.00000 18.279 <2e-16 ***
## scale(NO3_mgL_dl * 1000) -0.61361 0.29306 54.00000 -2.094 0.0410 *
## scale(NH4_mgL_dl * 1000) 0.28855 0.13960 54.00000 2.067 0.0435 *
## scale(PO4_ugL_dl) -0.03695 0.12569 54.00000 -0.294 0.7699
## scale(AFDM_mgg) -0.38478 0.18041 54.00000 -2.133 0.0375 *
## scale(wtr_m) 0.04233 0.22999 54.00000 0.184 0.8547
## scale(Q_m) 0.02593 0.13086 54.00000 0.198 0.8437
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(NO*1 s(NH*1 s(PO4_ s(AFDM scl(_)
## s(NO3_L_*10 -0.123
## s(NH4_L_*10 -0.157 -0.302
## scl(PO4_L_) -0.082 -0.035 -0.326
## scl(AFDM_m) -0.275 0.146 -0.310 0.055
## scal(wtr_m) -0.314 0.354 0.020 -0.100 -0.124
## scale(Q_m) -0.441 0.316 0.129 -0.141 0.165 0.387
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## scale(NO3_mgL_dl * 1000) scale(NH4_mgL_dl * 1000) scale(PO4_ugL_dl)
## 1.399768 1.468211 1.150803
## scale(AFDM_mgg) scale(wtr_m) scale(Q_m)
## 1.230784 1.346312 1.393157## R2m R2c
## [1,] 0.139549 0.139549B. Chlorophyll-a models
BW:
chla_mod_bw <- lmer(log(Chla_ugL_Q+1)~
scale(NO3_mgL_dl* 1000) +
scale(NH4_mgL_dl* 1000) +
scale(PO4_ugL_dl)+
scale(AFDM_mgg)+
#scale(DOC_mgL_dl)+
scale(wtr_m) +
scale(Q_m)+(1|site), data=covariat_datq_BW)
summary(chla_mod_bw)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Chla_ugL_Q + 1) ~ scale(NO3_mgL_dl * 1000) + scale(NH4_mgL_dl *
## 1000) + scale(PO4_ugL_dl) + scale(AFDM_mgg) + scale(wtr_m) +
## scale(Q_m) + (1 | site)
## Data: covariat_datq_BW
##
## REML criterion at convergence: 256.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.7932 -0.5506 -0.1391 0.4846 2.4774
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.4873 0.6981
## Residual 2.2105 1.4868
## Number of obs: 69, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.69172 0.54445 1.05614 4.944 0.1173
## scale(NO3_mgL_dl * 1000) 0.26748 0.18377 61.29299 1.456 0.1506
## scale(NH4_mgL_dl * 1000) -0.30309 0.16752 61.69467 -1.809 0.0753 .
## scale(PO4_ugL_dl) 0.14788 0.13070 61.25780 1.131 0.2623
## scale(AFDM_mgg) 0.20959 0.29054 61.36433 0.721 0.4734
## scale(wtr_m) 0.00381 0.31804 61.06972 0.012 0.9905
## scale(Q_m) -0.05790 0.18815 61.88681 -0.308 0.7593
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(NO*1 s(NH*1 s(PO4_ s(AFDM scl(_)
## s(NO3_L_*10 -0.132
## s(NH4_L_*10 0.035 -0.360
## scl(PO4_L_) 0.066 -0.163 -0.052
## scl(AFDM_m) 0.189 -0.262 0.076 0.118
## scal(wtr_m) -0.158 0.525 -0.220 -0.025 -0.228
## scale(Q_m) -0.199 0.434 0.021 -0.242 -0.370 0.618## scale(NO3_mgL_dl * 1000) scale(NH4_mgL_dl * 1000) scale(PO4_ugL_dl)
## 1.655305 1.245731 1.115461
## scale(AFDM_mgg) scale(wtr_m) scale(Q_m)
## 1.182510 1.998142 2.057583## R2m R2c
## [1,] 0.0826535 0.2483622GB:
chla_mod_gb <- lmer(log(Chla_ugL_Q+1)~
scale(NO3_mgL_dl* 1000) +
scale(NH4_mgL_dl* 1000) +
scale(PO4_ugL_dl)+
scale(AFDM_mgg)+
#scale(DOC_mgL_dl)+
scale(wtr_m) +
scale(Q_m) + (1|site), data=covariat_datq_GB)
summary(chla_mod_gb)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Chla_ugL_Q + 1) ~ scale(NO3_mgL_dl * 1000) + scale(NH4_mgL_dl *
## 1000) + scale(PO4_ugL_dl) + scale(AFDM_mgg) + scale(wtr_m) +
## scale(Q_m) + (1 | site)
## Data: covariat_datq_GB
##
## REML criterion at convergence: 220.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9286 -0.7348 0.1622 0.4396 2.1911
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.02052 0.1432
## Residual 0.72393 0.8508
## Number of obs: 83, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.44327 0.16017 1.08780 21.497 0.02305 *
## scale(NO3_mgL_dl * 1000) -0.35575 0.18547 75.95345 -1.918 0.05885 .
## scale(NH4_mgL_dl * 1000) 0.50245 0.15957 72.39039 3.149 0.00238 **
## scale(PO4_ugL_dl) -0.06480 0.08634 70.76286 -0.751 0.45538
## scale(AFDM_mgg) -0.15622 0.12522 39.22083 -1.248 0.21960
## scale(wtr_m) 0.34064 0.14431 64.56896 2.361 0.02128 *
## scale(Q_m) 0.17167 0.08714 75.96634 1.970 0.05248 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(NO*1 s(NH*1 s(PO4_ s(AFDM scl(_)
## s(NO3_L_*10 -0.238
## s(NH4_L_*10 0.050 -0.518
## scl(PO4_L_) -0.091 0.035 -0.424
## scl(AFDM_m) 0.060 0.324 -0.519 0.150
## scal(wtr_m) -0.450 0.465 -0.038 -0.147 0.006
## scale(Q_m) -0.317 0.355 0.008 -0.115 0.086 0.439## scale(NO3_mgL_dl * 1000) scale(NH4_mgL_dl * 1000) scale(PO4_ugL_dl)
## 2.059524 2.264356 1.306441
## scale(AFDM_mgg) scale(wtr_m) scale(Q_m)
## 1.406260 1.514430 1.336345## R2m R2c
## [1,] 0.2204337 0.2419218C. GPP models
BW:
gpp_mod_bw <- lmer(log(GPP_mean+1) ~
scale(NO3_mgL_dl* 1000) +
scale(NH4_mgL_dl* 1000) +
scale(PO4_ugL_dl)+
scale(AFDM_mgg)+
#scale(DOC_mgL_dl)+
scale(wtr_m) +
scale(Q_m)+ (1|site),
data=covariat_datq_BW_na)
summary(gpp_mod_bw)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(GPP_mean + 1) ~ scale(NO3_mgL_dl * 1000) + scale(NH4_mgL_dl *
## 1000) + scale(PO4_ugL_dl) + scale(AFDM_mgg) + scale(wtr_m) +
## scale(Q_m) + (1 | site)
## Data: covariat_datq_BW_na
##
## REML criterion at convergence: 74.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.3520 -1.0599 0.1177 0.6725 2.2217
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.0000 0.0000
## Residual 0.2085 0.4567
## Number of obs: 47, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.60126 0.07173 40.00000 8.382 2.41e-10 ***
## scale(NO3_mgL_dl * 1000) -0.03634 0.14040 40.00000 -0.259 0.797102
## scale(NH4_mgL_dl * 1000) 0.03251 0.10707 40.00000 0.304 0.762951
## scale(PO4_ugL_dl) 0.05057 0.07937 40.00000 0.637 0.527619
## scale(AFDM_mgg) 0.12434 0.08933 40.00000 1.392 0.171664
## scale(wtr_m) 0.47138 0.16981 40.00000 2.776 0.008331 **
## scale(Q_m) 0.41802 0.11104 40.00000 3.765 0.000536 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(NO*1 s(NH*1 s(PO4_ s(AFDM scl(_)
## s(NO3_L_*10 -0.012
## s(NH4_L_*10 -0.089 -0.791
## scl(PO4_L_) 0.128 -0.527 0.353
## scl(AFDM_m) 0.260 0.122 -0.269 0.086
## scal(wtr_m) -0.195 0.605 -0.477 -0.318 -0.028
## scale(Q_m) -0.105 0.125 -0.035 -0.013 -0.096 0.613
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## scale(NO3_mgL_dl * 1000) scale(NH4_mgL_dl * 1000) scale(PO4_ugL_dl)
## 3.943474 2.987120 1.453622
## scale(AFDM_mgg) scale(wtr_m) scale(Q_m)
## 1.151354 3.006878 1.951788## R2m R2c
## [1,] 0.3142777 0.3142777GB:
Highly skewed even when log transformed.
gpp_mod_gb <- lmer(log(GPP_mean+1) ~
scale(NO3_mgL_dl* 1000) +
scale(NH4_mgL_dl* 1000) +
scale(PO4_ugL_dl)+
scale(AFDM_mgg)+
#scale(DOC_mgL_dl)+
scale(wtr_m) +
scale(Q_m) + (1|site), data=covariat_datq_GB_na)
summary(gpp_mod_gb)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(GPP_mean + 1) ~ scale(NO3_mgL_dl * 1000) + scale(NH4_mgL_dl *
## 1000) + scale(PO4_ugL_dl) + scale(AFDM_mgg) + scale(wtr_m) +
## scale(Q_m) + (1 | site)
## Data: covariat_datq_GB_na
##
## REML criterion at convergence: -13.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.2910 -0.6469 -0.1149 0.2579 2.9371
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.0002591 0.0161
## Residual 0.0128017 0.1131
## Number of obs: 31, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.093556 0.033210 0.489448 2.817 0.3841
## scale(NO3_mgL_dl * 1000) 0.007607 0.025168 23.999945 0.302 0.7651
## scale(NH4_mgL_dl * 1000) 0.032072 0.020539 3.214033 1.562 0.2103
## scale(PO4_ugL_dl) -0.056735 0.029337 8.430856 -1.934 0.0873 .
## scale(AFDM_mgg) -0.003998 0.019357 8.228514 -0.207 0.8414
## scale(wtr_m) 0.009400 0.036439 23.006449 0.258 0.7987
## scale(Q_m) 0.004899 0.018976 23.055864 0.258 0.7986
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(NO*1 s(NH*1 s(PO4_ s(AFDM scl(_)
## s(NO3_L_*10 -0.044
## s(NH4_L_*10 0.141 -0.054
## scl(PO4_L_) -0.295 -0.549 -0.336
## scl(AFDM_m) -0.382 0.257 -0.289 0.045
## scal(wtr_m) -0.518 0.147 -0.079 -0.146 0.376
## scale(Q_m) -0.431 0.021 -0.067 0.133 0.220 0.334## scale(NO3_mgL_dl * 1000) scale(NH4_mgL_dl * 1000) scale(PO4_ugL_dl)
## 1.721912 1.307889 1.935489
## scale(AFDM_mgg) scale(wtr_m) scale(Q_m)
## 1.378244 1.340005 1.190482## R2m R2c
## [1,] 0.1675341 0.1840497fixed_effects_nep1 <- broom.mixed::tidy(gpp_mod_gb, effects = "fixed", conf.int = TRUE)
# Filter to exclude intercept and arrange predictors
fixed_effects_nep1 <- fixed_effects_nep1[fixed_effects_nep1$term != "(Intercept)", ]
fixed_effects_nep1 <- fixed_effects_nep1[order(fixed_effects_nep1$estimate), ] # Order by estimate
fixed_effects_nep1$Catchment <- "GB"D. |ER| models
BW:
er_mod_bw <- lmer(log(ab_ER+1) ~
scale(NO3_mgL_dl* 1000) +
scale(NH4_mgL_dl* 1000) +
scale(PO4_ugL_dl)+
scale(AFDM_mgg)+
#scale(DOC_mgL_dl)+
scale(wtr_m) +
scale(Q_m) + (1|site),
data=covariat_datq_BW_na)
summary(er_mod_bw)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(ab_ER + 1) ~ scale(NO3_mgL_dl * 1000) + scale(NH4_mgL_dl *
## 1000) + scale(PO4_ugL_dl) + scale(AFDM_mgg) + scale(wtr_m) +
## scale(Q_m) + (1 | site)
## Data: covariat_datq_BW_na
##
## REML criterion at convergence: -60.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.24094 -0.59081 0.01924 0.68879 3.15168
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.121445 0.34849
## Residual 0.006303 0.07939
## Number of obs: 47, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.565975 0.246736 0.995205 10.400 0.061662 .
## scale(NO3_mgL_dl * 1000) 0.014010 0.024947 39.016569 0.562 0.577598
## scale(NH4_mgL_dl * 1000) -0.012010 0.018634 39.000887 -0.645 0.523010
## scale(PO4_ugL_dl) -0.004434 0.015607 39.083279 -0.284 0.777816
## scale(AFDM_mgg) 0.002081 0.016103 39.027033 0.129 0.897859
## scale(wtr_m) 0.133738 0.034606 39.103309 3.865 0.000408 ***
## scale(Q_m) -0.133371 0.020169 39.032426 -6.613 7.3e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(NO*1 s(NH*1 s(PO4_ s(AFDM scl(_)
## s(NO3_L_*10 -0.001
## s(NH4_L_*10 -0.004 -0.783
## scl(PO4_L_) 0.004 -0.359 0.289
## scl(AFDM_m) 0.014 0.060 -0.247 -0.050
## scal(wtr_m) -0.010 0.613 -0.432 0.004 -0.161
## scale(Q_m) -0.006 0.177 -0.047 0.125 -0.165 0.652## scale(NO3_mgL_dl * 1000) scale(NH4_mgL_dl * 1000) scale(PO4_ugL_dl)
## 3.994356 2.919201 1.267362
## scale(AFDM_mgg) scale(wtr_m) scale(Q_m)
## 1.187484 3.437483 2.045603## R2m R2c
## [1,] 0.2068482 0.9608681# Extract fixed effects and confidence intervals
fixed_effects_er <- broom.mixed::tidy(er_mod_bw, effects = "fixed", conf.int = TRUE)
# Filter to exclude intercept and arrange predictors
fixed_effects_er <- fixed_effects_er[fixed_effects_er$term != "(Intercept)", ]
fixed_effects_er <- fixed_effects_er[order(fixed_effects_er$estimate), ] # Order by estimate
fixed_effects_er$Catchment <- "BW"GB:
er_mod_gb <- lmer(log(ab_ER+1) ~
scale(NO3_mgL_dl* 1000) +
scale(NH4_mgL_dl* 1000) +
scale(PO4_ugL_dl)+
scale(AFDM_mgg)+
# scale(DOC_mgL_dl)+
scale(wtr_m) +
scale(Q_m)+ (1|site),
data=covariat_datq_GB_na)
summary(er_mod_gb)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(ab_ER + 1) ~ scale(NO3_mgL_dl * 1000) + scale(NH4_mgL_dl *
## 1000) + scale(PO4_ugL_dl) + scale(AFDM_mgg) + scale(wtr_m) +
## scale(Q_m) + (1 | site)
## Data: covariat_datq_GB_na
##
## REML criterion at convergence: 30.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.97321 -0.72982 0.02548 0.61764 2.00362
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.66140 0.8133
## Residual 0.06701 0.2589
## Number of obs: 31, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.71037 0.58126 1.01003 2.943 0.20656
## scale(NO3_mgL_dl * 1000) -0.04198 0.05876 23.03354 -0.715 0.48210
## scale(NH4_mgL_dl * 1000) 0.11777 0.05831 23.27930 2.020 0.05507 .
## scale(PO4_ugL_dl) -0.07954 0.07675 23.19187 -1.036 0.31074
## scale(AFDM_mgg) 0.04697 0.05074 23.19413 0.926 0.36414
## scale(wtr_m) 0.27939 0.08671 23.06356 3.222 0.00376 **
## scale(Q_m) 0.28472 0.04344 23.00103 6.554 1.09e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(NO*1 s(NH*1 s(PO4_ s(AFDM scl(_)
## s(NO3_L_*10 -0.021
## s(NH4_L_*10 0.061 -0.160
## scl(PO4_L_) -0.072 -0.374 -0.524
## scl(AFDM_m) -0.082 0.317 -0.492 0.271
## scal(wtr_m) -0.087 0.193 -0.224 0.011 0.450
## scale(Q_m) -0.054 0.014 -0.033 0.099 0.175 0.312## scale(NO3_mgL_dl * 1000) scale(NH4_mgL_dl * 1000) scale(PO4_ugL_dl)
## 1.655592 1.912623 2.126215
## scale(AFDM_mgg) scale(wtr_m) scale(Q_m)
## 1.773162 1.395697 1.137209## R2m R2c
## [1,] 0.1538996 0.9221604fixed_effects_er1 <- broom.mixed::tidy(er_mod_gb, effects = "fixed", conf.int = TRUE)
# Filter to exclude intercept and arrange predictors
fixed_effects_er1 <- fixed_effects_er1[fixed_effects_er1$term != "(Intercept)", ]
fixed_effects_er1 <- fixed_effects_er1[order(fixed_effects_er1$estimate), ] # Order by estimate
fixed_effects_er1$Catchment <- "GB"Figure 3: General drivers for biomass, chl-a and GPP
(Q1) - How does stream metabolism vary based on catchment characteristics, antecedent flow conditions, and nitrogen availability?
Mainly addresses how productivity is associated with nutrients and water quality:
In general GPP may be more a function of physical conditions but is partially associated with NH4.
Part2/ Question 2
Nitrogen demand and supply dynamics
Calculate nitrogen supply and demand:
# Constants
ra <- 0.5 # Autotrophic respiration coefficient (Hall & Tank 2003)
C_Nauto <- 16 # Autotrophic C:N ratio (Stelzer & Lamberti 2001)
C_Nhetero <- 20 # Heterotrophic C:N ratio (Hall & Tank 2003)
HGE <- 0.05 # Heterotrophic Growth Efficiency (Hall & Tank 2003)
# Calculate components of nitrogen demand
covariat_nitrogen2 <- covariat_datq %>%
filter(substrate=="sw")%>% # select just surface water samples
group_by(site) %>%
arrange(date) %>% # Ensure chronological order
mutate( ## 17 NAs
v_mi = ifelse(is.na(v_m),
rollapplyr(v_m, width = 3, FUN = function(x) mean(x, na.rm = TRUE), fill = NA, partial = TRUE),
v_m),
w_mi = ifelse(is.na(w_m),
rollapplyr(w_m, width = 3, FUN = function(x) mean(x, na.rm = TRUE), fill = NA, partial = TRUE),
w_m)
) %>%
# Carry forward any remaining NAs
mutate(
v_mi = zoo::na.locf(v_mi, na.rm = FALSE),
v_mi = zoo::na.locf(v_mi, fromLast = TRUE, na.rm = FALSE),
w_mi = zoo::na.locf(w_mi, na.rm = FALSE),
w_mi = zoo::na.locf(w_mi, fromLast = TRUE, na.rm = FALSE)
) %>%
# If there are still missing values, replace them with the site mean
mutate(
v_mi = ifelse(is.na(v_mi), mean(v_m, na.rm = TRUE), v_mi),
w_mi = ifelse(is.na(w_mi), mean(w_m, na.rm = TRUE), w_mi)
) %>%
ungroup()%>%
## should already be at detection limits
# filter(water_year < 2024) %>%
# mutate(NO3_mgL_dl = if_else(NO3_mgL_dl < 0.0015, 0.0015, NO3_mgL_dl))%>%
# mutate(NH4_mgL_dl=if_else(NH4_mgL_dl < 0.001, 0.001, NH4_mgL_dl))%>%
# mutate(PO4_ugL_dl=if_else(PO4_ugL_dl < 0.201, 0.201, PO4_ugL_dl))%>%
mutate(
## ** Pause here to double check Q units
Q_Ls = Q_m * 1000, # flow from csm to Ls
# Autotrophic respiration (raGPP)
raGPP = GPP_mean * ra,
# Autotrophic assimilation of N
Auto_N_assim = GPP_mean / C_Nauto,
# Heterotrophic respiration (Rh)
Rh = ER_mean - raGPP,
# Heterotrophic assimilation of N
Hetero_N_assim = (Rh * HGE) / C_Nhetero,
# Total nitrogen demand
Ndemand = Auto_N_assim + Hetero_N_assim, # unit should be g N m-2 d-1
# calculate reach length in m
reachL = c(((v_mi*10)*w_mi*86.4)/K600_daily_mean),
Ndemand = if_else(Ndemand < 0, 0.0001, Ndemand), # have demand be essentially zero
# Calculate no3 supply
NO3_supply = c(((86400*Q_Ls*NO3_mgL_dl)/(w_mi*reachL))/1000),
# Calculate nh3 supply
NH4_supply = c(((86400*Q_Ls*NH4_mgL_dl)/(w_mi*reachL))/1000), ## unit should be g N m-2 d-1
PO4_supply = c(((86400*Q_Ls*(PO4_ugL_dl/1000))/(w_mi*reachL))/1000) ## unit should be g N m-2 d-1
)Plots of Ratios of NH4+- N supply to demand, ratios NO3– N supply to demand
Summary of reach based surface water supply
## [1] 4## [1] 7## [1] 38## # A tibble: 2 × 11
## catch N_supplym NO3_supplym NO3_supplymin NO3_supplymax NH4_supplym
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 BW 6.88 3.76 0.0526 213. 0.417
## 2 GB 18.5 9.13 0.554 73.6 6.14
## # ℹ 5 more variables: NH4_supplymin <dbl>, NH4_supplymax <dbl>,
## # PO4_supplym <dbl>, PO4_supplymin <dbl>, PO4_supplymax <dbl>Figure 4: Nitrogen supply and demand seasonally at each reach
(Q2) - Is variance in nitrogen demand and instream uptake (either as ammonium or nitrate) more related to biological (GPP, ER, biomass or chlorophyll-a), physical processes (flow and water temperature), or ambient nutrient and organic matter availability?
# n_s_grid <- ggarrange(
# NO3_supp_plots,
# N_demand_plot1,
# labels = c("b","c"),
# ncol = 1, nrow = 2)
n_s_grid <- ggarrange(
NO3_supp_plots,
N_demand_plot1,
labels = c("b", "c"),
ncol = 1, nrow = 2,
label.x = 0.95, # Move label to the right
label.y = 1, # Keep label at the top
hjust = 1, # Right-align the label
vjust = 1 # Top-align the label
)
#
# # Create the top row with extra spacing on both sides
# #left_col <- ggdraw() + draw_plot(N_ratio_plots, x = 0.15, height = 0.85) # Centered with 25% narrower width
# left_col <- ggdraw() +
# draw_plot(N_ratio_plots, x = 0.15, y = 0.075, height = 0.85)
### big grid
Ndyn_grid <- ggarrange(
N_ratio_plots,
n_s_grid,
ncol = 2, nrow = 1,
labels = c("a", ""),
common.legend = TRUE,
widths = c(1.75, 2.25),
legend = "bottom")
Ndyn_grid
Plots of Ratios of NH4+- N supply to demand, ratios NO3– N supply to demand colored by DOY
n_s_grid <- ggarrange(
NO3_supp_plots,
N_demand_plot1,
labels = c("b", "c"),
ncol = 1, nrow = 2,
label.x = 0.95, # Move label to the right
label.y = 1, # Keep label at the top
hjust = 1, # Right-align the label
vjust = 1 # Top-align the label
)
Ndyn_grid <- ggarrange(
N_ratio_plots,
n_s_grid,
ncol = 2, nrow = 1,
labels = c("a", ""),
common.legend = TRUE,
widths = c(1.75, 2.25),
legend = "bottom")
Ndyn_grid
Calculations of flow normalized NO3 and NH4 average daily loads by water year:
Plot for nitrogen cycling by reach location
Figure 5 a-c
Correlations between N uptake and …
- biomass
- chl-a
- GPP
- ER
- Q
- temp
Dataframe for successfull N uptatke
covariat_N <- covariat_datq %>%
filter(Uadd_ug_L_min < 121) %>%
filter(success=="yes")%>%
dplyr::select(site, catch, date, water_year, Chla_ugL_Q, method, Uadd_ug_L_min,
Uadd_min_sd, sw_m, sw_sd, Vf_m_min, vf_min_sd, GPP_mean, ER_mean,
AFDM_ugcm2, Q_m, wtr_m, AFDM_mgg) %>%
drop_na(Uadd_ug_L_min)
covariat_NU <- covariat_N %>% distinct()
#hist(covariat_NU$Uadd_ug_L_min, main = paste(NULL))
#hist(log(covariat_NU$Uadd_ug_L_min+1), main = paste(NULL))
covariat_NU_B <- covariat_NU%>%
filter(catch=="BW")
covariat_NU_G <- covariat_NU%>%
filter(catch=="GB")U_mod_OM <- lmer(log(Uadd_ug_L_min+1) ~
scale(AFDM_mgg) + (1|site), data=covariat_NU_B)
summary(U_mod_OM)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(AFDM_mgg) + (1 | site)
## Data: covariat_NU_B
##
## REML criterion at convergence: 49.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.3988 -0.5762 0.1545 0.4827 2.0762
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.6351 0.7969
## Residual 0.5798 0.7615
## Number of obs: 20, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.6466 0.6030 0.9346 4.389 0.1557
## scale(AFDM_mgg) -0.5344 0.1907 17.7468 -2.802 0.0119 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(AFDM_m) -0.087#hist(residuals(U_mod_OM),main = paste(NULL))
U_mod_bio_r2_vals <- r.squaredGLMM(U_mod_OM)
U_mod_bio_r2_vals## R2m R2c
## [1,] 0.1903139 0.613577U_mod_OM <- lmer(log(Uadd_ug_L_min+1) ~
scale(AFDM_mgg) + (1|site), data=covariat_NU_G)
summary(U_mod_OM)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(AFDM_mgg) + (1 | site)
## Data: covariat_NU_G
##
## REML criterion at convergence: 71.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.9161 -0.6175 -0.1110 0.3873 2.9432
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.0000 0.0000
## Residual 0.4754 0.6895
## Number of obs: 33, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.6164 0.1200 31.0000 21.799 <2e-16 ***
## scale(AFDM_mgg) -0.3295 0.1219 31.0000 -2.703 0.011 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(AFDM_m) 0.000
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')#hist(residuals(U_mod_OM),main = paste(NULL))
U_mod_bio_r2_vals <- r.squaredGLMM(U_mod_OM)
U_mod_bio_r2_vals## R2m R2c
## [1,] 0.1859186 0.1859186d) Uptake and epilithic biomass:
BW:
U_mod_bio <- lmer(log(Uadd_ug_L_min+1) ~
scale(AFDM_ugcm2) + (1|site), data=covariat_NU_B)
summary(U_mod_bio)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(AFDM_ugcm2) + (1 | site)
## Data: covariat_NU_B
##
## REML criterion at convergence: 34.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.70519 -0.73386 0.07471 0.87520 1.21436
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.5019 0.7085
## Residual 0.6167 0.7853
## Number of obs: 14, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.5939 0.5743 0.8498 4.517 0.1712
## scale(AFDM_ugcm2) -0.4012 0.2211 11.2340 -1.815 0.0964 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(AFDM_2) 0.056GB:
U_mod_biog <- lmer(log(Uadd_ug_L_min+1) ~
scale(AFDM_ugcm2) + (1|site), data=covariat_NU_G)
summary(U_mod_biog)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(AFDM_ugcm2) + (1 | site)
## Data: covariat_NU_G
##
## REML criterion at convergence: 56.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.55876 -0.69352 -0.08685 0.77110 2.09503
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.0000 0.0000
## Residual 0.7326 0.8559
## Number of obs: 22, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.46975 0.18249 20.00000 13.534 1.58e-11 ***
## scale(AFDM_ugcm2) -0.01063 0.18678 20.00000 -0.057 0.955
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(AFDM_2) 0.000
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')e) Uptake and chl-a:
BW:
U_mod_chla <- lmer(log(Uadd_ug_L_min+1) ~
scale(Chla_ugL_Q) + (1|site), data=covariat_NU_B)
summary(U_mod_chla)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(Chla_ugL_Q) + (1 | site)
## Data: covariat_NU_B
##
## REML criterion at convergence: 31.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.20993 -0.01623 0.27721 0.51286 1.33623
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.000 0.0000
## Residual 0.664 0.8148
## Number of obs: 13, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.5657 0.2260 11.0000 11.353 2.05e-07 ***
## scale(Chla_ugL_Q) -0.5248 0.2352 11.0000 -2.231 0.0474 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(Ch_L_Q) 0.000
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## R2m R2c
## [1,] 0.293197 0.293197GB:
U_mod_chlag <- lmer(log(Uadd_ug_L_min+1) ~
scale((Chla_ugL_Q)) + (1|site), data=covariat_NU_G)
summary(U_mod_chlag)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale((Chla_ugL_Q)) + (1 | site)
## Data: covariat_NU_G
##
## REML criterion at convergence: 32.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.0086 -0.5137 0.0005 0.7429 1.9490
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.0000 0.0000
## Residual 0.2141 0.4627
## Number of obs: 22, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.69228 0.09866 20.00000 27.290 < 2e-16 ***
## scale((Chla_ugL_Q)) -0.49926 0.10098 20.00000 -4.944 7.81e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## sc((C_L_Q)) 0.000
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## R2m R2c
## [1,] 0.5379097 0.5379097f) Uptake and GPP:
BW:
U_mod_GPP <- lmer(log(Uadd_ug_L_min+1) ~
scale(log(GPP_mean+1)) + (1|site), data=covariat_NU_B)
summary(U_mod_GPP)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(log(GPP_mean + 1)) + (1 | site)
## Data: covariat_NU_B
##
## REML criterion at convergence: 28.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.28637 -0.58032 -0.03618 0.21176 1.89708
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.09332 0.3055
## Residual 0.75517 0.8690
## Number of obs: 11, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.13774 0.33992 0.94015 6.289 0.111
## scale(log(GPP_mean + 1)) 0.05369 0.27707 8.37351 0.194 0.851
##
## Correlation of Fixed Effects:
## (Intr)
## s((GPP_+1)) 0.006GB:
U_mod_GPPg <- lmer(log(Uadd_ug_L_min+1) ~
scale(log(GPP_mean+1)) + (1|site), data=covariat_NU_G)
summary(U_mod_GPPg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(log(GPP_mean + 1)) + (1 | site)
## Data: covariat_NU_G
##
## REML criterion at convergence: 9.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.0848 -0.2151 0.1633 0.2847 0.8518
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 17.1989 4.1472
## Residual 0.1768 0.4205
## Number of obs: 5, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.9139 2.9418 0.9007 0.991 0.517
## scale(log(GPP_mean + 1)) 3.2949 0.7635 2.2249 4.315 0.041 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## s((GPP_+1)) 0.045g) Uptake and |ER|:
BW:
U_mod_ER <- lmer(log(Uadd_ug_L_min+1) ~
scale((ER_mean*-1)) + (1|site), data=covariat_NU_B)
summary(U_mod_ER)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale((ER_mean * -1)) + (1 | site)
## Data: covariat_NU_B
##
## REML criterion at convergence: 25.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.7221 -0.4105 0.1035 0.5490 1.3523
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.0000 0.0000
## Residual 0.5886 0.7672
## Number of obs: 11, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.1254 0.2313 9.0000 9.188 7.2e-06 ***
## scale((ER_mean * -1)) 0.4466 0.2426 9.0000 1.841 0.0988 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## s((ER_*-1)) 0.000
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')GB:
U_mod_ER <- lmer(log(Uadd_ug_L_min+1) ~
scale((ER_mean*-1)) + (1|site), data=covariat_NU_G)
summary(U_mod_ER)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale((ER_mean * -1)) + (1 | site)
## Data: covariat_NU_G
##
## REML criterion at convergence: 7.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.8134 -0.6034 -0.1037 0.3973 1.1232
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.09622 0.3102
## Residual 0.17679 0.4205
## Number of obs: 5, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.3118 0.2904 0.8941 7.960 0.0976 .
## scale((ER_mean * -1)) 0.9306 0.2155 2.1657 4.317 0.0430 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## s((ER_*-1)) -0.022h) Uptake and Q:
BW:
U_mod_Q <- lmer(log(Uadd_ug_L_min+1) ~
scale(log(Q_m+1)) + (1|site), data=covariat_NU_B)
summary(U_mod_Q)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(log(Q_m + 1)) + (1 | site)
## Data: covariat_NU_B
##
## REML criterion at convergence: 42.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.0014 -0.3472 -0.2333 0.5525 2.2220
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.1599 0.3999
## Residual 0.5333 0.7303
## Number of obs: 18, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.4174 0.3353 0.8893 7.209 0.1075
## scale(log(Q_m + 1)) 0.4530 0.1852 15.8680 2.446 0.0265 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl((Q_+1)) 0.047GB:
U_mod_Qg <- lmer(log(Uadd_ug_L_min+1) ~
scale(log(Q_m)+1) + (1|site), data=covariat_NU_B)
summary(U_mod_Qg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(log(Q_m) + 1) + (1 | site)
## Data: covariat_NU_B
##
## REML criterion at convergence: 44.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.9829 -0.5508 -0.1698 0.4880 2.2779
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.09962 0.3156
## Residual 0.61880 0.7866
## Number of obs: 18, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.3941 0.2946 0.8186 8.127 0.1116
## scale(log(Q_m) + 1) 0.3448 0.1974 15.9941 1.747 0.0998 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl((Q_)+1) 0.044i) Uptake and temp:
BW:
U_mod_temp <- lmer(log(Uadd_ug_L_min+1) ~
scale(wtr_m) + (1|site), data=covariat_NU_B)
summary(U_mod_temp)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(wtr_m) + (1 | site)
## Data: covariat_NU_B
##
## REML criterion at convergence: 45
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.4477 -0.6101 -0.2310 0.5474 2.1467
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.0000 0.0000
## Residual 0.6829 0.8264
## Number of obs: 18, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.3398 0.1948 16.0000 12.013 2.03e-09 ***
## scale(wtr_m) -0.2563 0.2004 16.0000 -1.279 0.219
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scal(wtr_m) 0.000
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## R2m R2c
## [1,] 0.08772759 0.08772759GB:
U_mod_tempg <- lmer(log(Uadd_ug_L_min+1) ~
scale(wtr_m) + (1|site), data=covariat_NU_G)
summary(U_mod_tempg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(wtr_m) + (1 | site)
## Data: covariat_NU_G
##
## REML criterion at convergence: 71.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.0146 -0.6432 0.1866 0.6857 2.3690
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.05039 0.2245
## Residual 0.63804 0.7988
## Number of obs: 29, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.5496 0.2175 0.9197 11.720 0.0652 .
## scale(wtr_m) 0.1399 0.1540 26.9996 0.908 0.3718
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scal(wtr_m) 0.010## R2m R2c
## [1,] 0.02764357 0.09881207U_mod_NO3 <- lmer(log(Uadd_ug_L_min+1) ~
scale(NO3_mgL_dl) + (1|site), data=covariat_datq_BW)
summary(U_mod_NO3)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(NO3_mgL_dl) + (1 | site)
## Data: covariat_datq_BW
##
## REML criterion at convergence: 183.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.29637 -0.73973 0.05334 0.99156 2.17629
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.0000 0.0000
## Residual 0.5996 0.7743
## Number of obs: 77, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.24906 0.08956 75.00000 25.113 <2e-16 ***
## scale(NO3_mgL_dl) -0.11925 0.07567 75.00000 -1.576 0.119
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(NO3_L_) -0.171
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## R2m R2c
## [1,] 0.03164349 0.03164349U_mod_NO3 <- lmer(log(Uadd_ug_L_min+1) ~
scale(NO3_mgL_dl) + (1|site), data=covariat_datq_GB%>%filter(NO3_mgL_dl<0.7780))
summary(U_mod_NO3)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(NO3_mgL_dl) + (1 | site)
## Data: covariat_datq_GB %>% filter(NO3_mgL_dl < 0.778)
##
## REML criterion at convergence: 169.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.2861 -0.2404 0.1208 0.4928 2.2231
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.1917 0.4378
## Residual 0.8532 0.9237
## Number of obs: 62, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.52938 0.33122 0.99863 7.637 0.0831 .
## scale(NO3_mgL_dl) 0.05608 0.24008 59.10558 0.234 0.8161
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(NO3_L_) -0.019## R2m R2c
## [1,] 0.0007345274 0.1840798U_mod_NH4 <- lmer(log(Uadd_ug_L_min+1) ~
scale(NH4_mgL_dl) + (1|site), data=covariat_datq_BW%>%filter(substrate=="pw"))
summary(U_mod_NH4)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(NH4_mgL_dl) + (1 | site)
## Data: covariat_datq_BW %>% filter(substrate == "pw")
##
## REML criterion at convergence: 65.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.9969 -0.8189 -0.1895 0.7163 2.2768
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.000 0.0000
## Residual 0.515 0.7177
## Number of obs: 29, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.2989 0.1335 27.0000 17.219 4.37e-16 ***
## scale(NH4_mgL_dl) 0.3134 0.1040 27.0000 3.013 0.00556 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(NH4_L_) -0.061
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## R2m R2c
## [1,] 0.244845 0.244845U_mod_NH4 <- lmer(log(Uadd_ug_L_min+1) ~
scale(NH4_mgL_dl) + (1|site), data=covariat_datq_GB%>%filter(substrate=="pw"))
summary(U_mod_NH4)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Uadd_ug_L_min + 1) ~ scale(NH4_mgL_dl) + (1 | site)
## Data: covariat_datq_GB %>% filter(substrate == "pw")
##
## REML criterion at convergence: 94.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.9824 -0.4760 -0.0800 0.4451 2.9994
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.06356 0.2521
## Residual 0.42799 0.6542
## Number of obs: 45, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.65248 0.20399 0.91437 13.003 0.0600 .
## scale(NH4_mgL_dl) -0.26077 0.09683 42.35027 -2.693 0.0101 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(NH4_L_) 0.021## R2m R2c
## [1,] 0.1394342 0.2507134Plots of Ut and all linear model covariates by stream
Figure 5: General drivers of either NO3 or NH4 uptake rates (Ut)
Metabolism by water year:
Dataframes:
For metabolism by water year March - October (2021-2024)
Metabolimsm regression analysis :
For nitrogen dynamics by water year March - October (2021-2023)
## [1] 0.755418## [1] 0.205832## [1] 0.6## [1] 0.4347887Linear models for metabolism across dry or wet years
lmm_WY_GPP <- lmer(GPP_mean ~ WY_lab + (1|site) + (1|catch), data = metab_WY_lmdf)
summary(lmm_WY_GPP)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GPP_mean ~ WY_lab + (1 | site) + (1 | catch)
## Data: metab_WY_lmdf
##
## REML criterion at convergence: 5150.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.1777 -0.5184 -0.1201 0.4356 6.6409
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.06227 0.2495
## catch (Intercept) 1.96383 1.4014
## Residual 1.67037 1.2924
## Number of obs: 1531, groups: site, 4; catch, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.77100 1.00001 1.00257 1.771 0.327
## WY_labnormal -0.92626 0.08342 1526.89813 -11.104 <2e-16 ***
## WY_labwet -1.47328 0.08208 1509.10270 -17.950 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) WY_lbn
## WY_labnorml -0.028
## WY_labwet -0.032 0.327## R2m R2c
## [1,] 0.09936061 0.5930161lmm_WY_GPP_BW <- lmer(GPP_mean ~ WY_lab + (1|site), data = metab_WY_lmdf%>%filter(catch=="BW"))
summary(lmm_WY_GPP_BW)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GPP_mean ~ WY_lab + (1 | site)
## Data: metab_WY_lmdf %>% filter(catch == "BW")
##
## REML criterion at convergence: 2807.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.0656 -0.6096 -0.1146 0.4305 4.6548
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.1986 0.4456
## Residual 2.8763 1.6960
## Number of obs: 719, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.3460 0.3287 1.0789 10.18 0.0528 .
## WY_labnormal -1.9112 0.1812 715.3341 -10.55 <2e-16 ***
## WY_labwet -2.6190 0.1460 714.0794 -17.94 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) WY_lbn
## WY_labnorml -0.129
## WY_labwet -0.189 0.285## R2m R2c
## [1,] 0.3257379 0.3692876lmm_WY_GPP_GB <- lmer(GPP_mean ~ WY_lab + (1|site), data = metab_WY_lmdf%>%filter(catch=="GB"))
summary(lmm_WY_GPP_GB)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GPP_mean ~ WY_lab + (1 | site)
## Data: metab_WY_lmdf %>% filter(catch == "GB")
##
## REML criterion at convergence: 172.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.7543 -0.4954 -0.1098 -0.0996 11.0680
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.00000 0.0000
## Residual 0.07097 0.2664
## Number of obs: 812, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.20156 0.01352 809.00000 14.903 < 2e-16 ***
## WY_labnormal -0.17131 0.02153 809.00000 -7.957 5.92e-15 ***
## WY_labwet -0.11845 0.02445 809.00000 -4.844 1.53e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) WY_lbn
## WY_labnorml -0.628
## WY_labwet -0.553 0.347
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## R2m R2c
## [1,] 0.077609 0.077609lmm_WY_ER <- lmer((ER_mean*-1) ~ WY_lab + (1|site) + (1|catch), data = metab_WY_lmdf)
summary(lmm_WY_ER)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: (ER_mean * -1) ~ WY_lab + (1 | site) + (1 | catch)
## Data: metab_WY_lmdf
##
## REML criterion at convergence: 7703.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.1922 -0.4857 0.0530 0.5767 4.1823
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 21.673 4.655
## catch (Intercept) 17.285 4.158
## Residual 8.827 2.971
## Number of obs: 1531, groups: site, 4; catch, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 10.1707 3.7516 1.0008 2.711 0.225
## WY_labnormal -0.2992 0.1918 1525.0800 -1.560 0.119
## WY_labwet -2.2281 0.1892 1525.2642 -11.779 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) WY_lbn
## WY_labnorml -0.017
## WY_labwet -0.019 0.327## R2m R2c
## [1,] 0.01872544 0.8187402lmm_WY_ER_BW <- lmer((ER_mean*-1) ~ WY_lab + (1|site), data = metab_WY_lmdf%>%filter(catch=="BW"))
summary(lmm_WY_ER_BW)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: (ER_mean * -1) ~ WY_lab + (1 | site)
## Data: metab_WY_lmdf %>% filter(catch == "BW")
##
## REML criterion at convergence: 3802.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9074 -0.6550 0.1540 0.6621 3.0292
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 23.64 4.862
## Residual 11.49 3.390
## Number of obs: 719, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 15.0178 3.4430 1.0028 4.362 0.14295
## WY_labnormal -1.1740 0.3622 715.0129 -3.241 0.00124 **
## WY_labwet -4.8123 0.2923 715.1599 -16.466 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) WY_lbn
## WY_labnorml -0.025
## WY_labwet -0.036 0.284## R2m R2c
## [1,] 0.119547 0.7119466lmm_WY_ER_GB <- lmer((ER_mean*-1) ~ WY_lab + (1|site), data = metab_WY_lmdf%>%filter(catch=="GB"))
summary(lmm_WY_ER_GB)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: (ER_mean * -1) ~ WY_lab + (1 | site)
## Data: metab_WY_lmdf %>% filter(catch == "GB")
##
## REML criterion at convergence: 3449.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0389 -0.4967 -0.2066 0.5091 5.0202
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 20.045 4.477
## Residual 4.039 2.010
## Number of obs: 812, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.3550 3.1678 1.0013 1.690 0.34
## WY_labnormal 0.6775 0.1644 808.0233 4.121 4.15e-05 ***
## WY_labwet 0.8754 0.1899 808.0543 4.610 4.67e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) WY_lbn
## WY_labnorml -0.022
## WY_labwet -0.021 0.370## R2m R2c
## [1,] 0.006115623 0.8333196lmm_WY_NO3_Q <- lmer(NO3_supply ~ WY_lab + (1|site) + (1|catch),, data = covariat_WY)
summary(lmm_WY_NO3_Q)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: NO3_supply ~ WY_lab + (1 | site) + (1 | catch)
## Data: covariat_WY
##
## REML criterion at convergence: 591.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.4708 -0.3057 -0.2494 -0.0662 6.9382
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 24.53 4.953
## catch (Intercept) 0.00 0.000
## Residual 826.57 28.750
## Number of obs: 64, groups: site, 4; catch, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.288 5.769 4.971 1.090 0.326
## WY_labnormal -4.460 20.908 59.542 -0.213 0.832
## WY_labwet 6.042 7.780 35.558 0.777 0.442
##
## Correlation of Fixed Effects:
## (Intr) WY_lbn
## WY_labnorml -0.191
## WY_labwet -0.610 0.143
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## R2m R2c
## [1,] 0.01204887 0.04052549lmm_WY_NH4_Q <- lmer(NH4_supply ~ WY_lab + (1|site) + (1|catch), data = covariat_WY)
summary(lmm_WY_NH4_Q)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: NH4_supply ~ WY_lab + (1 | site) + (1 | catch)
## Data: covariat_WY
##
## REML criterion at convergence: 352.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8007 -0.4836 -0.2467 0.2983 3.1005
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 31.23 5.588
## catch (Intercept) 0.00 0.000
## Residual 44.64 6.681
## Number of obs: 54, groups: site, 4; catch, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.4901 3.1800 3.8527 0.154 0.8852
## WY_labnormal -3.0047 4.9126 48.1412 -0.612 0.5437
## WY_labwet 4.9876 1.9978 49.7590 2.497 0.0159 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) WY_lbn
## WY_labnorml -0.097
## WY_labwet -0.350 0.153
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## R2m R2c
## [1,] 0.08667957 0.4625832lmm_WY_demand_Q <- lmer(Ndemand ~ WY_lab + (1|site) + (1|catch), data = covariat_WY)
summary(lmm_WY_demand_Q)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Ndemand ~ WY_lab + (1 | site) + (1 | catch)
## Data: covariat_WY
##
## REML criterion at convergence: -218.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.3101 -0.5609 -0.2249 0.4222 3.0156
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.0003462 0.01861
## catch (Intercept) 0.0027512 0.05245
## Residual 0.0029978 0.05475
## Number of obs: 81, groups: site, 4; catch, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.08040 0.03940 1.05094 2.040 0.2805
## WY_labnormal -0.06722 0.03968 75.26618 -1.694 0.0944 .
## WY_labwet -0.07771 0.01326 70.13310 -5.858 1.38e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) WY_lbn
## WY_labnorml -0.042
## WY_labwet -0.169 0.120## R2m R2c
## [1,] 0.1983456 0.6057187plots
Figure 6: Average annual differences in GPP, ER, N-demand and N-supply by water year
Values averaged across March to October observations
Extra?
Relationship between Ut and n-demand? not really there.
Correlations btwn GPP and other standing stock estiamtes of productivity
Looking at stoichiometeric ratios of N:P overtime?
N:P
Looks like N:P <16 across all observations
Statisical relationship?
Maybe? GPP and N:P are slightly negatively related. But falls apart when both catchments are modeled independently
mod <- lmer((GPP_mean) ~ scale(N_P_ratio) + (1|catch), data = Stoich_df %>%filter(water_year<2024))
summary(mod)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: (GPP_mean) ~ scale(N_P_ratio) + (1 | catch)
## Data: Stoich_df %>% filter(water_year < 2024)
##
## REML criterion at convergence: 163.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.4738 -0.4973 -0.1762 0.0920 2.7231
##
## Random effects:
## Groups Name Variance Std.Dev.
## catch (Intercept) 1.121 1.059
## Residual 1.551 1.246
## Number of obs: 49, groups: catch, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.7355 0.7719 1.0053 0.953 0.5147
## scale(N_P_ratio) -0.5650 0.2958 46.0356 -1.910 0.0623 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(N_P_rt) 0.058## R2m R2c
## [1,] 0.04253083 0.444168modb <- lmer((GPP_mean) ~ scale(N_P_ratio) + (1|site), data = Stoich_df %>%filter(water_year<2024 & catch=="BW"))
summary(modb)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: (GPP_mean) ~ scale(N_P_ratio) + (1 | site)
## Data: Stoich_df %>% filter(water_year < 2024 & catch == "BW")
##
## REML criterion at convergence: 110.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.3352 -0.5563 -0.3577 0.5690 2.0395
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.2365 0.4863
## Residual 2.3756 1.5413
## Number of obs: 30, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.3925 0.4515 0.9108 3.084 0.219
## scale(N_P_ratio) -0.7080 0.5118 27.4652 -1.383 0.178
##
## Correlation of Fixed Effects:
## (Intr)
## scl(N_P_rt) 0.148## R2m R2c
## [1,] 0.06103181 0.1460509modg <- lmer((GPP_mean) ~ scale(N_P_ratio) + (1|site), data = Stoich_df %>%filter(water_year<2024 & catch=="GB"))
summary(modg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: (GPP_mean) ~ scale(N_P_ratio) + (1 | site)
## Data: Stoich_df %>% filter(water_year < 2024 & catch == "GB")
##
## REML criterion at convergence: -16.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.6729 -0.5975 -0.4147 0.3845 3.2583
##
## Random effects:
## Groups Name Variance Std.Dev.
## site (Intercept) 0.00000 0.000
## Residual 0.01664 0.129
## Number of obs: 19, groups: site, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.07495 0.03047 17.00000 2.459 0.0249 *
## scale(N_P_ratio) 0.02788 0.04747 17.00000 0.587 0.5647
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(N_P_rt) 0.238
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')## R2m R2c
## [1,] 0.01880352 0.01880352
Statisical relationship?
No
mod1 <- lmer((ER_mean*-1) ~ scale(N_P_ratio) + (1|catch), data = Stoich_df %>%filter(water_year<2024))
summary(mod1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: (ER_mean * -1) ~ scale(N_P_ratio) + (1 | catch)
## Data: Stoich_df %>% filter(water_year < 2024)
##
## REML criterion at convergence: 253.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8286 -0.6234 -0.2085 0.8830 2.5149
##
## Random effects:
## Groups Name Variance Std.Dev.
## catch (Intercept) 45.06 6.713
## Residual 10.06 3.171
## Number of obs: 49, groups: catch, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 8.2966 4.7709 1.0010 1.739 0.332
## scale(N_P_ratio) 0.4253 0.7531 46.0061 0.565 0.575
##
## Correlation of Fixed Effects:
## (Intr)
## scl(N_P_rt) 0.024## R2m R2c
## [1,] 0.00121901 0.8177674Table of mean N:P ratios by site
## # A tibble: 4 × 6
## site Tot_N_uMm N_P_ratiom N_P_ratiomin N_P_ratiomax rangeNP
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 BWL 0.559 1.60 0.306 4.11 3.81
## 2 BWU 0.642 3.43 0.223 13.3 13.1
## 3 GBL 0.805 1.76 0.175 4.91 4.73
## 4 GBU 0.827 1.76 0.332 6.14 5.81Moddified figure 6?
(3) How does in-stream productivity and nitrogen supply and demand dynamics change within stream catchments and across dry (2021 and 2022) and wet water years (2023)?
